Vol. 18: 17–25, 2012 ENDANGERED SPECIES RESEARCH Published online July 20 doi: 10.3354/esr00431 Endang Species Res

Succession and disturbance in an endangered red spruce−Fraser forest in the southern , , USA

Philip B. White, Saskia L. van de Gevel*, Peter T. Soulé

Department of Geography and Planning, Appalachian Ring Lab, Appalachian State University, Boone, North Carolina 28608, USA

ABSTRACT: Red spruce− forests are geographically limited to high elevations in the Appalachian Mountains (USA) and are considered to be endangered in the USA. We investigated the successional status and radial growth patterns in the heavily disturbed red spruce Sarg. and Fraser fir Abies fraseri (Pursh) Poir. forest of Roan Mountain, and North Carolina. This study elucidates the complexity of second-growth red spruce development after logging and disturbances by piceae Ratz. We documented precise temporal information of stand age, disturbance regimes, recruitment patterns, and the successional trajectory of the spruce−fir forest community. We used radial growth patterns of red spruce samples to detect the frequency and magnitude of disturbance. Red spruce was the oldest dominant canopy species, although Fraser fir had high recruitment rates over the past 80 yr. Changes in forest structure and species richness coincided with stand-wide disturbance events such as balsam woolly adelgid infestation and widespread early twentieth-century logging. The competitive advantage of Fraser fir over red spruce has resulted in an even-aged Fraser fir-domi- nant forest that occupies a relatively early stage of successional development. This study provides a 130 yr environmental history to assist land managers in the southern Appalachian Mountains as they develop long-term restoration plans for this unique ecosystem.

KEY WORDS: Disturbance · Spruce−fir forest · Dendroecology · Stand dynamics · Logging · Balsam woolly adelgid

Resale or republication not permitted without written consent of the publisher

INTRODUCTION 2002, Rentch et al. 2007), as anthropogenic activities have altered disturbance vectors and successional pat- The red spruce Picea rubens Sarg. and Fraser fir terns. The forest community exists in only a few dis- Abies fraseri (Pursh) Poir. forests of the southern junct, island populations at high elevations in the Appalachian Mountains are ranked as the second most southern Appalachian Mountains. The RSFF forest has endangered ecosystem in the USA (Noss et al. 1995, greater diversity than similar northern forests Christensen et al. 1996, Rentch et al. 2007). Within the (Delcourt & Delcourt 2000), and sustains several en - eastern USA, these subalpine forests have undergone demic, endangered species such as the Carolina north- a major shift in structure and composition during the ern flying squirrel Glaucomys sabrinus coloratus, past several decades (Johnson & Siccama 1983, Cook spruce-fir moss spider Micohexura montivaga, Roan et al. 1987, Dull et al. 1988, Eagar & Adams 1992, Goelz Mountain bluet Hedyotis purpurea var. montana, and et al. 1999, Busing 2004, Rentch et al. 2007, 2010). rock gnome lichen Gymnoderma lineare (Rentch et al. There is concern about the re sil ience and vigor of the 2010). This forest type is highly valued by ecologists, red spruce−Fraser fir (RSFF) ecosystem (Wear & Greis park and land managers, and the public.

*Corresponding author. Email: [email protected] © Inter-Research 2012 · www.int-res.com 18 Endang Species Res 18: 17–25, 2012

During the mid-20th century, scientists observed ern Appalachian Mountains, where endemic en - widespread growth decline and tree mortality in dangered species are present. In the present study, RSFF forests (Hornbeck & Smith 1985, Adams & we quantified the spruce−fir forest age, the current Eagar 1992). The balsam woolly adelgid Adelges forest composition, and the successional trajectory of piceae Ratz., an exotic that kills mature Fraser the spruce−fir forest on Roan Mountain in the south- fir , has cyclically infested southern Ap pa la - ern Appalachian Mountains. The primary objectives chian Mountain Fraser fir forests since the 1950s, were to (1) quantitatively document the current com- affecting most RSFF forests (Boyce & Martin 1993, position and structure of Roan Mountain’s RSFF for- M. Pyne & D. Durham unpubl. data). Mortality rates est, (2) use dendroecological techniques to investi- for mature Fraser fir trees are high (Dull et al. 1988), gate the forest’s disturbance history and to elucidate and the species is becoming in creasingly less domi- its pre-logging composition, and (3) determine how nant (Busing et al. 1993, Busing 2004). land use and disturbance history have affected the Human disturbance during the 19th and 20th cen- successional development of this forest. turies has affected the forest health and successional trajectory of southern Appalachian Mountain forests. Spruce−fir forests were altered by the aggressive log- MATERIALS AND METHODS ging practices, burning, and subsequent soil erosion which were common in the region during the late Study area 1800s and early 1900s (Pyle 1984, McLaughlin et al. 1991). The great economic value of red spruce Roan Mountain is a part of the Unaka Mountains, a lumber during the early 1900s led to heavy logging subset of the southern Appalachian Mountains and with little long-term management planning (Korstian part of the Blue Ridge physiographic province (Clark 1937, Pyle 1984, Hayes et al. 2007). RSFF forests were 2008). The mountain is located at approximately selectively logged or clear-cut. Extensively logged 36° 6’ 16.42’’ N, 82° 7’ 47.39’’ W, its ridgeline delineat- spruce−fir forests include the Great Smoky Mountain ing the state border between Tennessee and North National Park of Tennessee and North Carolina, the Carolina (Fig. 1). Roan Mountain (including Roan Balsam Mountains, the Black Mountains, Grand - High Bluff and Roan High Knob) encompasses father Mountain, the Plott Balsams, and Roan Moun- ap prox imately 19 km2. The entire Roan Massif, tain in North Carolina; and Mount Rogers, Virginia however, is much larger and extends along the (Pyle & Schafale 1988, Smith & Nicholas 1999, Hayes North Carolina− Tennessee border for approximately et al. 2007). Pyle (1984) found that as much as 50% of 30 km. Much of the massif is jointly managed by Pis- all Appa lach ian spruce−fir forests were replaced by gah National Forest in North Carolina and Cherokee hardwood species after logging. National Forest in Tennessee. Roan Mountain’s high Among the highest peaks of the southern Appala - elevations are classified as Cfb (marine temperate chian Mountains, Roan Mountain hosts one of the climate) under the Köppen climate classification sys- few remaining RSFF forest communities. Located at tem (Christopherson 2006), based on the cooler tem- the boundary between Tennessee and North Car- peratures and increased precipitation associated olina, the Roan Massif area and its spruce−fir forests with high elevation and orographic uplift (Christo- are over 1650 m elevation. Logging began on Roan pherson 2006). Soils are primarily well-drained Mountain during the late 1800s, but was minimal rel- Inceptisol loams that form on steep, rocky slopes and ative to the aggressive logging during the 1920s and ridge tops (NRCS 2010). 1930s when the area was clear-cut, removing trees The Roan Mountain spruce−fir forest was surveyed larger than 15 cm in diameter (Wilson 1991). Logging prior to 1930s logging (Brown 1941) and RSFF were lasted until 1937 and effectively denuded the highest found to represent 89.2% of all trees (62.3% Fraser elevations of Roan Mountain (Wilson 1991). In 1941, fir and 26.9% red spruce). Yellow birch Betula the USDA Forest Service purchased approximately alleghaniensis Britt., mountain maple Acer spicatum 2800 ha atop Roan Mountain, granting protection to Lam., American beech Fagus grandifolia Ehrh., the area’s endangered spruce−fir forest (Wilson mountain ash Sorbus americana Marsh., yellow 1991). Logging has not occurred on Roan Mountain buck eye Aesculus octandra Aiton, and pin cherry since the USDA Forest Service acquired the land. Prunus pensylvanica L. were also present. Although Previous studies have focused on spruce−fir forests Fraser fir had the highest density, red spruce had the in the central Appalachian Mountains, but few stud- highest basal area. Fraser fir was the most common ies have examined this forest ecosystem in the south- tree in the understory (Brown 1941). White et al.: Red spruce−Fraser fir succession 19

Field methods into 3 size classes (Class 1: <2.5 cm DBH; Class 2: 2.5−4.9 cm DBH; Class 3: 5.0−9.9 cm DBH). We established 6 circular 0.05 ha fixed-radius plots (r = 12.66 m) within RSFF codominant stands to study forest composition, canopy structure, and forest dis- Laboratory methods turbance patterns. All plots were located approxi- mately 1800 m or higher in uneven-aged stands. We Increment cores were processed following stan- targeted stands with a variety of age classes to dard dendroecological techniques (Stokes & Smiley develop a historical timeline of disturbance events. 1996). We dated and crossdated all tree cores. Most Plots were separated by at least 100 m and were cores intersected or very nearly intersected the tree’s downslope from the generally southwest to northeast pith. We calculated density, basal area (dominance), trend of the ridgeline. We measured the location, and importance values of each tree species (Cottam percent slope, and aspect of each plot. We collected & Curtis 1956, Ludwig & Reynolds 1988, Matthews & increment cores from all trees ≥10 cm diameter at Mackie 2007, Hart et al. 2008). Importance values breast height (DBH, 1.37 m above ground) to deter- were calculated as the sum of the relative density mine tree age. We cored trees close to the ground to and relative dominance. obtain the maximum amount of growth rings (Fritts We measured annual growth rings from red spruce 1976). Cores with extensive rot and decay were dis- cores to the nearest 0.001 mm with a Velmex measur- carded, but these were relatively few in number. ing system coupled with Measure J2X software. The We inventoried tree species and canopy classes of crossdating accuracy of each red spruce series was trees ≥10 cm DBH. We categorized canopy position then statistically verified against all other red spruce based on the amount and direction of intercepted series using the computer program COFECHA and light and designated each tree as occupying either 30 yr segments lagged successively by 15 yr (Holmes dominant, codominant, intermediate, or suppressed 1983, Grissino-Mayer 2001). positions (Oliver & Larson 2002). In each overstory We standardized all series in the red spruce chro- plot, we established a fixed-radius 0.01 ha (r = 5.66) nology to remove effects from age-related growth nested sub-plot to evaluate regeneration of tree spe- trends that could add noise to the series unrelated to cies in the understory. All saplings (woody stems the climate signal desired in chro no logy de velop- <10 cm DBH, ≥1 m height) were tallied by species ment (Fritts 1976). We removed the age-related growth trend of each sample using the program ARSTAN (Cook 1985), which fits a negative exponential trend line to the growth of the sample using the least squares technique. ARSTAN then creates an index for that year by dividing the actual ring width by the value predicted by the regression (Fritts 1976, Cook 1985). The indices were then averaged for each year across all series to create a single red spruce chrono logy (Fritts 1976). The detection of release events in trees is a common approach for infer- ring the disturbance history of a forest community (Lorimer & Frelich 1989, Rubino & McCarthy 2004, Hart et al. 2008). A release event is defined as a period in which radial growth ab ruptly increases over a period of successive years (Nowacki & Abrams 1997, Rubino & McCarthy 2004). Release Fig. 1. Picea rubens and Abies fraseri. Red spruce and Fraser fir high-elevation forest islands in the southern Appalachian Mountains, USA. The present study events occur following a canopy dis- was conducted on Roan Mountain in North Carolina turbance (Fraver & White 2005). We 20 Endang Species Res 18: 17–25, 2012

used the computer program JOLTS to Table 1. Density, dominance, and relative importance values (i.e. mean of rel- identify release events in red spruce ative density + relative dominance) of trees located in the spruce-fir forest at tree-ring measurement series (Hol - Roan Mountain, North Carolina: red spruce Picea rubens, Fraser fir Abies fraseri, yellow birch , mountain ash Sorbus americana, mes 1999). The JOLTS program ana- and pin cherry Prunus pensylvanica. BA: basal area (m2) lyzes raw ring-width measurement values to identify release events Species Density Relative Dominance Relative Relative based upon user-defined parameters. (trees ha−1) density (BA ha−1) dominance importance We analyzed changes in ring width using a 10 yr running average of the Abies fraseri 810 83.22 38.22 80.06 81.63 previous and subsequent 10 yr. We Picea rubens 150 15.41 9.32 19.52 17.47 Sorbus 6.67 0.69 0.10 0.21 0.45 specified release detection para - americana meters of 100% for a major release, Prunus 3.33 0.34 0.07 0.14 0.24 50% for a moderate release, and 25% pensylvanica for a minor release, with a minimum Betula 3.33 0.34 0.03 0.07 0.20 allegheniensis of 5 yr between each release event. Total 973.33 100 47.74 100 100 For example, a major release event had a ring width 2.0 times greater than the average of the preceding and following 10 yr. A release event was defined as a stand-wide disturbance if 25% or more of all trees experienced a release in a given year (Nowacki & Abrams 1997, Rubino & McCarthy 2004). These synchronous release events were indicative of a large-scale exogenous disturbance, while asynchronous events (fewer than 25%) were considered localized and related to single tree deaths (Orwig & Abrams 1994).

RESULTS

Stand composition and age structure Fig. 2. Canopy positions of tree species surveyed in The dominant species above 1800 m at Roan Moun- spruce−fir forest plots at Roan Mountain, North Carolina: red spruce Picea rubens, Fraser fir Abies fraseri; other spe- tain was Fraser fir, accounting for 83.21% of all trees, cies include yellow birch Betula alleghaniensis, mountain with the highest relative dominance and importance maple Acer spicatum, American beech Fagus grandifolia, scores (Table 1). The other common species was red mountain ash Sorbus americana, yellow buckeye Aesculus spruce, and together, Fraser fir and red spruce com- octandra, and pin cherry Prunus pensylvanica prised 98.62% of the trees, with combined relative dominance and importance values of 99.58% and Age structure analysis revealed a distinct shift in 99.1% for spruce and fir, respectively. Hardwood forest composition during the late 1930s and early species present included mountain ash, pin cherry, 1940s, coinciding with the cessation of logging activ- and yellow birch (Table 1). Fraser fir was also the ities (Fig. 3). This period marked a change from a red dominant species in every canopy class (Fig. 2). Most spruce-dominated forest to a densely populated trees were in the codominant canopy position (of Fraser fir forest with fewer red spruce trees. A few which 83.8% were Fraser fir and 15.6% were red red spruce trees established prior to intensive log- spruce). The high density of Fraser fir compared to ging. The oldest tree in the stand was a red spruce red spruce remained consistent among all other that was at least 136 yr old, and the oldest Fraser fir canopy classes. In the understory, Fraser fir accounted was 109 yr old. Of all red spruce trees, 55.56% had for 85.92% of all saplings, dominating all size classes pith dates before 1937, which coincided with the final (Table 2). We calculated 4734 saplings ha−1 from the year of logging operations, while 82.77% of all Fraser sapling size classes. Mountain ash was the second fir trees sampled had pith dates of 1937 or later. most frequent species in the sapling survey. Red Although the post-logging pulse in tree establish- spruce was not a common understory species. ment was found for Fraser fir and red spruce trees, White et al.: Red spruce−Fraser fir succession 21

Table 2. Abies fraseri, Picea rubens, and Sorbus americana. ted Fraser from further release analysis because Sapling summary statistics for Fraser fir, red spruce, and the trees exhibited erratic juvenile growth and were mountain ash (Class 1 = <2.5 cm diameter at breast height, less than 80 yr old. We analyzed 26 red spruce series DBH; Class 2 = 2.5−4.9 cm DBH; Class 3 = 5.0−9.9 cm DBH) for radial growth changes and release events (Figs. 4 & 5). Of these, 23 trees (88%) experienced 91 release Species Class Total Sapling Relative 1 2 3 density density events that were at least minor (25% increase; (ind. ha−1) ha−1 Table 3). Most red spruce experienced multiple release events, with a mean of 3.96 release events Abies fraseri 127 99 18 244 4067.48 85.92 per tree. Fewer red spruce trees experienced moder- Picea rubens 7 8 1 16 266.72 5.63 Sorbus 23 1 0 24 400.8 8.45 ate or major events. We found 62 moderate release americana events in 21 trees and 29 major release events in 13 Total 157 108 19 284 4734.28 100 trees. Stand-wide release events occurred periodi- cally and only at minor and moderate release levels (Fig. 4). We found minor and moderate stand-wide events in 1935, 1940, and 2000. The most widespread release events occurred during the 1930s and 1940s as a result of logging activities. Radial growth releases related to logging lasted for over 2 decades and gradually diminished during the 1950s. The greatest stand-wide event occurred in 2000, when 25% of trees experienced moderate releases and 50% of trees experienced minor releases.

DISCUSSION

The RSFF forest at Roan Mountain experienced significant changes during the 20th century. Red spruce and Fraser fir shared dominance in the pre- logging forest, but the more rapidly regenerating Fraser fir was the dominant tree in the post-logging forest. Fraser fir benefitted from clear-cut logging Fig. 3. Picea rubens and Abies fraseri. (A) Age−diameter re- lationships of 286 red spruce and Fraser fir trees at Roan during the 1920s and 1930s, as the establishment Mountain, North Carolina. The black vertical line indicates of new Fraser firs dramatically outnumbered red the cessation of logging and subsequent pulse in tree estab- spruce establishment during the post-logging period. lishment. (B) Decade of establishment of the red spruce and Although we detected a spruce−fir establishment Fraser fir trees. DBH: diameter at breast height cohort post-logging, Fraser fir overwhelmingly dom- inated this establishment pulse, and red spruce was Fraser fir density was higher. Among all trees sam- largely absent from subsequent pulses in recruit- pled, 91.17%, of which 90.78% were Fraser fir, had ment. Fraser fir dominance over red spruce following post-logging pith dates. Following the post-logging disturbance is expected, as red spruce is shade toler- establishment period, tree establishment gradually ant (Busing et al. 1988, Smith & Nicholas 2000, Bus- decreased until a second, smaller establishment ing 2004). The modern Fraser fir-dominated forest pulse occurred during the 1970s and 1980s. The sec- was largely an even-aged stand that established dur- ond tree establishment pulse was mostly Fraser firs. ing the late 1930s and 1940s. Older individuals that were not logged likely exhibited irregular growth forms or were too small, making them undesirable to Disturbance history loggers. The contrast between the pre-logging and post-logging forest compositions suggests that clear- Most Fraser firs established post-logging. All cutting at Roan Mountain’s high elevations dramati- Fraser fir trees exhibited prolonged radial growth cally altered the spruce−fir forest. characteristics that reflect release episodes during The relative dominance, importance values, and the period immediately following logging. We omit- understory survey suggest that the modern high- 22 Endang Species Res 18: 17–25, 2012

single, horizontal crown stratum typi- cal of single-cohort stands. Trees growing as co dominants, with few dominant trees emerging from the canopy, are spatially limited horizon- tally, as crowns are crowded from all sides by other trees. This forest condi- tion is characteristic of a stand experi- encing the stem exclusion and under- story initiation stages of development (Oliver & Larson 2002). As the Roan Mountain spruce−fir forest ages, Fraser fir dominance may slightly de - crease as more shade-tolerant red spruce stems slowly reach the canopy. Fraser fir’s dominant status will con- tinue as the stand approaches old growth stage, albeit to a lesser extent, because fir typically occurs more fre- quently than red spruce above 1800 m in the southern Appalachian Moun- tains (Delcourt & Delcourt 2000). In the absence of future major disturbances, Fraser fir will likely continue its do- minance, although minor canopy dist - urbances may allow red spruce to in - crease in dominance. As the stand matures and aging Fraser fir trees become more susceptible to balsam woolly adelgid infestation, older Fraser firs will likely die. If the balsam woolly adelgid outbreak occurred with the successional decline of Fraser fir, red spruce could eventually become co - dominant, or in an extreme case, even dominant. Fig. 4. Picea rubens. (A) Minor, (B) moderate, and (C) major release events in red spruce at Roan Mountain, North Carolina, using a 10 yr running mean. The disturbance history of the Roan Vertical black bars represent the percentage of trees that experienced a re- Mountain spruce−fir forest is tied to lease in that year. Horizontal black lines represent sample size. Releases in its logging history, and all distur- 25% or more of the living trees were considered a stand-wide release bances that have occurred during the short existence of the current forest elevation forest at Roan Mountain currently occupies should be considered in the context of its logging an early to mid-successional stage (Oliver & Larson past. The stand-age and release event results 2002). Further, the high density of Fraser fir com- demonstrate the effect logging had on forest devel- pared to red spruce indicates that the forest is in a opment. Of the 3 minor and moderate stand-wide relatively early stage of development, because Fraser release events we detected, 2 (1935 and 1940) fir is a pioneer species (Messier et al. 1999). Fraser fir occurred after the cessation of logging. Localized also dominates the understory classes, indicating that disturbances occurred often during the post-logging insufficient time has passed for the shade tolerance era on both minor and moderate scales, while major of red spruce to confer a competitive advantage over releases occurred less frequently. These distur- Fraser fir. bances could be caused by the balsam woolly adel- The prevalence of the codominant canopy class gid, which first appeared on Roan Mountain in suggests that the trees in this stand have formed a 1962. Our analysis detected a major release event White et al.: Red spruce−Fraser fir succession 23

1.5 ever, the only documentation of Fraser fir dieback Picea rubens from the balsam woolly adelgid occurred during 1962, and our release event results do not show a clear balsam woolly adelgid infestation signal. It is difficult to assign these localized events solely to pulses in balsam woolly adelgid infestation, because thunderstorms, severe winds, and ice storms may 1.0 cause a similar disturbance signal. The stand-wide Index disturbance event during 2000 may have been related to a weather event, but we suspect it was related to a southern pine beetle Dendroctonus front - alis Zimmermann infestation that affected red spruce on Roan Mountain (Billings 2011). Although the 0.5 southern pine beetle does not commonly infest red 1860 1880 1900 1920 1940 1960 1980 2000 spruce, there have been some observed incidents in Year the southeastern USA during warm and dry condi- Fig. 5. Picea rubens. Ring-width index chronology gener- tions (Thatcher et al. 1980). ated from red spruce series at Roan Mountain, North Car- During the 1980s, scientists were concerned about olina (1874−2009). The mean radial growth is standardized the decline in red spruce health. Research was con- to 1.0. Note periods of radial growth releases during the 1870s, 1890s, 1930s, 1980s, 1990s, and 2000s ducted to investigate how climatic factors limit growth of red spruce across its range (Cook et al. 1987, Eagar & Adams 1992). Previous studies found Table 3. Picea rubens. Red spruce trees exhibiting minor, moderate, and major releases, total releases found in all red spruce to be a useful indicator of climate change trees, and mean number of releases per tree (Cook 1988), but little research on the species has taken place in the southern Appalachian Mountains Release Series with Releases in Mean no. since the 1980s, and the relationships between tree intensity releases all trees releases tree−1 growth, climate, and stand dynamics in the southern Appalachian spruce−fir forest are not fully under- Minor 23 91 3.96 stood (Cook & Zedaker 1992, Rentch et al. 2012). Moderate 21 62 2.95 Major 13 29 2.23 Future research related to climate change in RSFF forests should investigate how shifts in the climate regime affected RSFF forest productivity during that year as well as a slight increase in minor and recent decades. moderate release events from 1962 to the 1970s. None of these events, however, were stand-wide in nature, indicating that the ef fects of this initial in- CONCLUSIONS fest ation were perhaps not as severe at Roan Moun- tain as those documented in other areas of the Roan Mountain spruce−fir stands established be - southern Appalachians (Dull et al. 1988). We hypo- tween the 1880s and the early 1900s and are thesize that the relatively young age of this Fraser currently in the understory reinitiation stage. The fir forest during the initial infestation limited the stand-wide logging disturbance during the 1930s effects of the adelgid because juvenile Fraser firs drama tically altered the composition and structure of are more resistant to infestation than mature trees the area’s RSFF forest. The forest was clear-cut and is (Potter et al. 2005). Therefore, the devastating log- relatively even-aged. Tree-ring evidence from the ging of the 1930s may have mitigated the initial RSFF forest indicated great differences between the effects of the balsam woolly adelgid on the stand historic and modern forests. Red spruce was formerly dynamics of Roan Mountain’s spruce−fir forest over of greater importance in the RSFF community at the past 80 yr. Roan Mountain than it is currently. Red spruce tree A pattern of localized minor, moderate, and major rings showed high initial growth rates consistent release events during the mid-1980s, early 1990s, with the stand initiation stage. The red spruce that and 2000 indicates a broadly decadal cycle of distur- survived logging during the 1930s showed that trees bance episodes. These events were likely the result were greater in size and occupied upper canopy of major balsam woolly adelgid infestations. How- positions during the pre-logging period. The pre-log- 24 Endang Species Res 18: 17–25, 2012

ging forest also exhibited greater variation in age Boyce SG, Martin WH (1993) The future of the terrestrial classes than the current forest. A cohort of Fraser fir communities of the southeastern coastal plain. In: Martin WH, Boyce SG, Echternacht AC (eds) Biodiversity of the trees that established following logging dominated southeastern United States: upland terrestrial communi- the spruce−fir forest in the modern, post-logging era. ties. Wiley, New York, NY, p 339−366 The earlier successional advantage of Fraser fir over Brown DM (1941) Vegetation of Roan Mountain: a phyto- red spruce has resulted in an even-aged Fraser fir- sociological and successional study. Ecol Monogr 11: 61−97 dominant forest that occupies a relatively early stage Busing RT (2004) Red spruce dynamics in an old southern of successional development. In the absence of future Appalachian forest. J Torrey Bot Soc 131: 337−342 stand-wide disturbances, dominance of Fraser fir will Busing RT, Clebsch EEC, Eagar C, Pauley EF (1988) Two continue. decades of change in a spruce- RSFF forests are much less extensive today than fir forest. Bull Torrey Bot Club 115: 25−31 Busing RT, White PS, MacKenzie MD (1993) Gradient analy- during the late 19th century, and this forest type is sis of old spruce-fir forests of the Great Smoky Mountains now considered to be endangered (Noss et al. 1995). circa 1935. Can J Bot 71:951−958 Forest management options in the central Appa la - Christensen NL, Bartuska AM, Brown JH, Carpenter S and chian Mountains have been addressed in recent others (1996) The report of the Ecological Society of America Committee on the scientific basis for ecosystem studies (e.g. Hornbeck & Kochenderfer 1998, Schuler management. Ecol Appl 6:665−691 et al. 2002, Rentch et al. 2007, 2010, Adams et al. Christopherson RW (2006) Geosystems: an introduction to 2010). Forest thinning is a possible management physical geography, 5th edn. Pearson Prentice Hall, option to promote red spruce abundance in the cen- Upper Saddle River, NJ tral Appalachian Mountains (Rentch et al. 2010). Clark SHB (2008) Geology of the Southern Appalachian Mountains: U.S. Geological Survey Scientific Investiga- However, management plans would need to be mod- tions Map 2830. Available at http: // pubs.usgs. gov/ sim/ ified for the southern Appalachian Mountains. Log- 2830/ (accessed 11 May 2010) ging disturbance would negatively impact associated Cook ER (1985) A time series analysis approach to tree-ring endangered species endemic to the southern Appa- standardization. PhD dissertation, University of Arizona, Tucson, AZ la chian Mountain spruce−fir ecosystem. Research on Cook ER (1988) A tree ring analysis of red spruce in the spruce−fir stand dynamics and endangered species southern Appalachian Mountains. In: Van Deusen PC in the southern Appalachian Mountains could use (ed) Analyses of Great Smoky Mountain red spruce tree this study as baseline data representing spruce−fir ring data. USDA Forest Service Gen Tech Rep SO-69. Southern Forest Experiment Station, New Orleans, LA, forest conditions over the past 130 yr. Thus, this study p 6−19 provides a foundation for land managers in the Cook ER, Zedaker SM (1992) The dendroecology of red southern Appalachian Mountains to develop long- spruce decline. In: Eagar C, Adams MB (eds) Ecology term red spruce restoration plans. and decline of red spruce in the eastern United States. Springer Verlag, New York, NY, p 192−231 Cook ER, Johnson AH, Blasing TJ (1987) Forest decline: Acknowledgements. This study was supported by a re - modeling the effect of climate in tree rings. Tree Physiol search grant from the North Carolina Native Plant Society, 3: 27−40 as well as graduate student awards from Appalachian State Cottam G, Curtis JT (1956) The use of distance measures in University. A. Cochran provided invaluable cartographic phytosociological sampling. Ecology 37: 451−460 support. K. White, J. Holcolm, L. Bowman, W. Gandy, and B. Delcourt HR, Delcourt PA (2000) Eastern deciduous forests. White assisted with spruce−fir data collection. The Pisgah In: Barbour MG, Billings WD (eds) North American ter- and Cherokee National Forests graciously provided re - restrial vegetation, 2nd edn. Cambridge University search permits. Press, Cambridge, p 357−395 Dull CW, Ward JD, Brown HD, Ryan GW, Clerke WH, Uhler RJ (1988) Evaluation of spruce and fir mortality in the LITERATURE CITED Southern Appalachian Mountains. Protection Report R8. USDA Forest Service, Southern Region, Atlanta, GA Adams MB, Eagar C (1992) Impacts of acidic deposition on Eagar C, Adams MB (1992) The ecology and decline of red high-elevation spruce-fir forests: results from the Spruce- spruce in the eastern United States. Springer-Verlag, Fir Research Cooperative. For Ecol Manag 51:195−205 New York, NY Adams HS, Stephenson SL, Rollins AW, Adams MB (2010) Fraver S, White AS (2005) Identifying growth releases in The isolated red spruce communities of Virginia and West dendrochronological studies of forest disturbance. Can J Virgina. Proc Conf Ecol Manag High-Elevation Forests in For Res 35: 1648−1656 the Central and Southern Appalachian Mountains. GTR- Fritts HC (1976) Tree rings and climate. Academic Press, NRS-P-64. USDA Forest Service, Asheville, NC, p 1−12 New York, NY Billings RF (2011) Use of chemicals for prevention and con- Goelz JCG, Burk TE, Zedaker SM (1999) Long-term growth trol of southern pine beetle infestations. In: Coulson RN, trends of red spruce and Fraser fir at Mt. Rogers, Virginia Klepzig KD (eds) Southern pine beetle II. GTR-SRS-140. and Mt. Mitchell, North Carolina. For Ecol Manag 115: USDA Forest Service, Asheville, NC, p 367−379 49−59 White et al.: Red spruce−Fraser fir succession 25

Grissino-Mayer HD (2001) Evaluating crossdating accuracy: app/ WebSoilSurvey.aspx (accessed 11 May 2010) a manual and tutorial for the computer program Oliver CD, Larson BC (2002) Forest stand dynamics. COFECHA. Tree-Ring Res 57:205–221 McGraw-Hill, New York, NY Hart JL, van de Gevel SL, Grissino-Mayer HD (2008) Land Orwig DA, Abrams MD (1994) Land-use history (1720− use and forest dynamics in a natural area of the southern 1992), composition, and dynamics of oak-pine forests ridge and valley, Tennessee. Nat Areas J 28: 275−289 within the Piedmont and Coastal Plain of northern Vir- Hayes M, Moody A, White PS, Costanza JL (2007) The influ- ginia. Can J For Res 24:2141−2149 ence of logging and topography on the distribution of Potter KM, Frampton J, Sidebottom JR (2005) Impacts of bal- spruce-fir forests near their southern limits in Great sam woolly adelgid on the southern Appalachian spruce- Smoky Mountains National Park, USA. Plant Ecol 189: fir ecosystem and on the North Carolina Christmas tree 59−70 industry. In: Proc 3rd Symp in Holmes RL (1983) Computer assisted quality control in tree- the eastern United States, Asheville, NC, p 104−114 ring dating and measurement. Tree-Ring Bull 43: 69−78 Pyle C (1984) Pre-park disturbance in the spruce-fir forests Holmes RL (1999) Dendrochronology program library and of Great Smoky Mountains National Park. In: White PS the dendroecology program library. Laboratory of Tree- (ed) The southern Appalachian spruce-fir ecosystem: its Ring Research, University of Arizona, Tucson, AZ www. biology and threats. Research/Resource Management ltrr. arizona.edu/pub/dpl-mac/68k/dpl.txt Report SER-71. USDI National Park Service, Atlanta, Hornbeck JW, Kochenderfer JN (1998) Growth trends and GA management implications for West Virginia’s red spruce Pyle C, Schafale MP (1988) Land use history of three spruce- forests. North J Appl For 15:197−202 fir forest sites in the southern Appalachians. J For Hist Hornbeck JW, Smith RB (1985) Documentation of red spruce 32: 4−21 growth decline. Can J For Res 15: 1199−1201 Rentch JS, Schuler TM, Ford WM, Nowacki GJ (2007) Red Johnson AH, Siccama TG (1983) Acid deposition and forest spruce stand dynamics, simulations, and restoration decline. Environ Sci Technol 17:294A–305A opportunities in the Central Appalachians. Restor Ecol Korstian CF (1937) Perpetuation of spruce on cut-over and 15: 440−452 burned lands in the higher southern Appalachian Moun- Rentch JS, Schuler TM, Nowacki GJ, Beane NR, Ford WM tains. Ecol Monogr 7:125−167 (2010) Canopy gap dynamics of second-growth red Lorimer CG, Frelich LE (1989) A methodology for estimating spruce-northern hardwood stands in West Virginia. For canopy disturbance frequency and intensity in dense Ecol Manag 260: 1921−1929 temperate forests. Can J For Res 19: 651−663 Rubino DL, McCarthy BC (2004) Comparative analysis of Ludwig JA, Reynolds JF (1988) Statistical ecology. J. Wiley, dendroecological methods used to assess disturbance New York, NY events. Dendrochronologia 21:97−115 Matthews RW, Mackie ED (2007) Forest mensuration: a Schuler TM, Ford WM, Collins RJ (2002) Successional handbook for practitioners. Forestry Commission, Edin- dynamics and restoration implications of a montane burgh coniferous forest in the central Appalachians. Nat Areas McLaughlin SB, Andersen CP, Hanson PJ, Tjoelker MG, Roy J 22: 88−98 WK (1991) Increased dark respiration and calcium defi- Smith GF, Nicholas NS (1999) Post-disturbance spruce-fir ciency of red spruce in relation to acidic deposition at forest stand dynamics at seven disjunct sites. Castanea high-elevation southern Appalachian Mountain sites. 64: 175−186 Can J For Res 21: 1234−1244 Smith GF, Nicholas NS (2000) Size- and age-class distribu- Messier C, Doucet R, Ruel JC, Claveau Y, Kelly C, Lechow- tions of Fraser fir following balsam woolly adelgid infes- icz MJ (1999) Functional ecology of advance regenera- tation. Can J For Res 30: 948−957 tion in relation to light in boreal forests. Can J For Res 29: Stokes MA, Smiley TL (1996) An introduction to tree-ring 812−823 dating. University of Arizona Press, Tucson, AZ Noss RF, LaRoe ET, Scott JM (1995) Endangered ecosystems Thatcher RC, Searcy JL, Coster JE, Hertel GD (1980) The of the United States: a preliminary assessment of loss and southern pine beetle. Tech Bull 1631. USDA Forest Ser- degradation. Biol Rep 28. USDI National Biological Ser- vice, Expanded Southern Pine Beetle Research and vice, , DC Application Program, Science and Education Adminis- Nowacki GJ, Abrams MD (1997) Radial-growth averaging tration, Pineville, LA criteria for reconstructing disturbance histories from pre- Wear DN, Greis JG (2002) Southern forest resource assess- settlement origin oaks. Ecol Monogr 67:225−249 ment: summary of findings. J For 100: 6−15 NRCS (Natural Resource Conservation Service) (2010) Soil Wilson JB (1991) Roan Mountain: a passage of time. Blair, survey. Available at http:// websoilsurvey. nrcs. usda.gov/ Winston-Salem, NC

Editorial responsibility: Hans Juergen Boehmer, Submitted: December 2, 2011; Accepted: March 6, 2012 Bonn, Germany Proofs received from author(s): June 20, 2012